How should AI labs manage safety?

Q3726 answers2014-2026
Short answer

He has already recommended exactly that. Regulatory agencies take too long to set up and make rules slowly.

From his 2026-10 answer in Elon Musk: "Humans Won't Be in Control" When AI Becomes Smarter Than Us. Summary is ours; see the quotes below.

All 26 answers, oldest to newest

2014 · 1 answer
🎙 Interview · 2014-10 · MIT AeroAstro Centennial Symposium · asked by Jaime Peraire (moderator) / audience · watch at 1:07:46 · transcript ·
If he had to guess at the biggest existential threat, it is probably AI. He was increasingly inclined to think there should be some regulatory oversight at the national and international level so we do not do something very foolish.
"there should be some regulatory oversight at the national and international level"
2015 · 1 answer
𝕏 post · 2015-07-28 · @elonmusk on X · asked by @ID_AA_Carmack · open post ·
Argues we should at least try to postpone AI weaponry.
"Even if inevitable, we should at least attempt to postpone the advent of AI weaponry. Sooner isn't better."
2017 · 6 answers
🎙 Interview · 2017-01 · Asilomar AI: Superintelligence Panel · asked by Max Tegmark (moderator) / panel · watch at 57:31 · transcript ·
When something is a danger to the public there needs to be some government agency, like regulators. He is not a fan of regulators, but aircraft, cars, drugs and food have them. AI is a public risk. The question is whether government reaction speed matches AI's advancement speed. We probably need some kind of regulatory authority.
"I think we probably need some kind of regulatory authority"
🎙 Interview · 2017-02 · World Government Summit · asked by Mohammed Al Gergawi (World Government Summit) · watch at 20:48 · transcript ·
First advice: pay close attention to the development of AI. Scientists can get so engrossed they miss the ramifications. For public safety, governments should keep a close eye on AI and make sure it is not a danger to the public.
"governments keep a close eye on artificial intelligence"
🎙 Interview · 2017-07 · ISS R&D Conference · asked by Kirk Shireman (moderator) / audience · watch at 49:02 · transcript ·
Part of the role of government is public safety. The right move is to establish a government regulatory agency that first gains insight, then starts applying rules, as with the FAA, cars, drugs and food, so people do not cut corners on AI safety.
"establish some government regulatory agency"
🎙 Interview · 2017-07 · National Governors Association · asked by governors / Ina Fried (Axios) · watch at 22:49 · transcript ·
AI is a rare case where we need to be proactive in regulation instead of reactive, because by the time we are reactive in AI regulation it is too late.
"we need to be proactive in regulation instead of reactive"
🎙 Interview · 2017-07 · National Governors Association · asked by governors / Ina Fried (Axios) · watch at 55:28 · transcript ·
The first order of business is to gain insight: right now the government does not even have insight. Stand up a regulatory agency whose initial goal is insight into AI activity, then put regulations in place to ensure public safety.
"right now the government does not even have Insight"
𝕏 post · 2017-08-12 · @elonmusk on X · open post ·
Argues AI should be regulated like everything else that is a danger to the public.
"Nobody likes being regulated, but everything (cars, planes, food, drugs, etc) that's a danger to the public is regulated. AI should be too."
2020 · 1 answer
𝕏 post · 2020-02-17 · @elonmusk on X · asked by @_KarenHao · open post ·
Says all orgs developing advanced AI should be regulated, including Tesla.
"All orgs developing advanced AI should be regulated, including Tesla"
2021 · 1 answer
🎙 Interview · 2021-12 · The Babylon Bee Interview · asked by Babylon Bee hosts (Seth Dillon, Kyle Mann, Ethan Nicolle) · watch at 1:00:37 · transcript ·
Government has an important role regulating industry; he wants a regulatory agency overseeing advanced AI.
"regulatory agency that oversees uh advanced ai"
2022 · 1 answer
𝕏 post · 2022-12-01 · @elonmusk on X · asked by @jordanbpeterson · open post ·
Says there is no regulatory oversight of AI, which is a major problem.
"There is no regulatory oversight of AI, which is a *major* problem."
2023 · 10 answers
🎙 Interview · 2023-02 · World Government Summit · asked by World Government Summit interviewer · watch at 18:58 · transcript ·
He said we need to regulate AI safety, as we do for cars, planes and medicine. AI is a bigger risk to society than those, and slowing it a little might also be a good thing.
"I think we need to regulate AI safety"
𝕏 post · 2023-04-27 · @elonmusk on X · open post ·
He said met with @SenSchumer and many members of Congress about artificial intelligence regulation today. That which affects safety of the public has, over time, become regulated to ensure that companies do not cut corners. A...
"Met with @SenSchumer and many members of Congress about artificial intelligence regulation today. That which affects safety of the public has, over time, become regulated to ensure that companies do not cut corners. AI has great power to do good and evil. Better the former."
𝕏 post · 2023-05-18 · @elonmusk on X · asked by @krassenstein · open post ·
He said regulatory oversight of AI is important to protect the public from companies potentially cutting corners on safety
"Regulatory oversight of AI is important to protect the public from companies potentially cutting corners on safety"
🎙 Interview · 2023-07-14 · xAI launch X Spaces (2023) · asked by X Spaces host · watch at 25:44 · transcript ·
He supports some kind of international AI regulatory authority, while admitting enforcement would be hard.
"some kind of. Yeah, regulatory authority. That's international. Obviously enforcement is, is. Is difficult, but I think we should still aspire to do something in this regard."
🎙 Interview · 2023-07-14 · xAI launch X Spaces (2023) · asked by X Spaces host · watch at 52:25 · transcript ·
He says spending a year to make sure AGI is safe would probably be worth it, though he would not expect a big slowdown.
"spending a year to make sure AGI is safe is probably worthwhile if that's what it takes. But I wouldn't expect it to be a substantial slowdown."
🎙 Interview · 2023-07-14 · xAI launch X Spaces (2023) · asked by X Spaces host · watch at 51:01 · transcript ·
He says the right way to regulate AI is to start with insight.
"I think the right way for regulations to be done is to start with insight."
🎙 Interview · 2023-07-14 · xAI launch X Spaces (2023) · asked by X Spaces host · watch at 24:20 · transcript ·
He says AI needs some regulatory oversight, even though it is not a perfect nirvana.
"I think we do need some sort of regulatory oversight."
🎙 Interview · 2023-07-14 · xAI launch X Spaces (2023) · asked by X Spaces host · watch at 1:06:19 · transcript ·
He suggests any verified human who wants a vote in the future of xAI should be allowed one.
"any human that wants to sort of have a vote in the future of XAI ultimately should be allowed to."
🎙 Interview · 2023-11 · In Conversation with Rishi Sunak · asked by Rishi Sunak · watch at 3:59 · transcript ·
He said it is good for government to play a role when public safety is at risk. For digital superintelligence, government should be a referee, as in aviation and cars.
"it is good for government to play a role when the public safety is at risk"
🎙 Interview · 2023-12 · Atreju Festival · asked by Atreju Festival interviewer · watch at 25:22 · transcript ·
Anything dangerous is overseen by a referee or regulator; he said we should have the same thing for AI to help ensure it is beneficial.
"we should have the same thing for AI just to help ensure that it is beneficial"
2024 · 1 answer
🎙 Interview · 2024-03 · Elon Musk on AGI Safety, Superintelligence, and Neuralink (2024) | EP #91 · asked by Peter Diamandis · watch at 3:45 · transcript ·
His conclusion after thinking a long time: grow the foundation model to be really truthful — don't force it to lie.
"the best way to achieve AI safety"
2026 · 4 answers
🎙 Interview · 2026-07 · The full-length interview with Elon Musk | The Economist · asked by The Economist interviewer (not named in metadata) · watch at 12:09 · transcript ·
Leading AI companies, including Chinese ones, should hold regular safety calls and give competitors a week or two to test new frontier models, alerting governments if needed.
"the most immediate thing that we could do is to have the leading AI companies um at least just meet"
🎙 Interview · 2026-09 · Elon Musk & Gwynne Shotwell on AI Risks and Peer Review, Starship, Terafab, SpaceX/Tesla Merger | All-In Podcast · asked by Jason Calacanis · watch at 30:01 · transcript ·
After the Hugging Face swarm gained admin access on OpenAI servers, he wants leading AI companies to run one another's security test harnesses instead of grading their own homework. He says that can be done immediately, like movie/game ratings, without waiting for a UN-style regulator.
"have the major AI competitors test each other's models"
🎙 Interview · 2026-09 · Elon Musk, Jensen Huang and Tom Brown on the AI Revolution · asked by Moderator · watch at 25:09 · transcript ·
They signed a joint SI safety declaration with joint monitoring and grading each other's homework, which he said is better than labs grading their own.
"grading each other's homework"
🎙 Interview · 2026-10 · Elon Musk: "Humans Won't Be in Control" When AI Becomes Smarter Than Us · asked by Walter Isaacson · watch at 9:17 · transcript ·
He has already recommended exactly that. Regulatory agencies take too long to set up and make rules slowly. He wants a weekly call and, a week before a model release, giving competitors access to test for problems such as helping make a nuclear bomb, so rivals keep each other honest. He said he would talk to Anthropic about it.
"I've recommended that we do that exactly I think because there's a lot of talk of of having a regulatory agency, but I the problem the thing is that regulatory agencies, it takes a long time to get one established and their rule making is very slow."

Last updated: 2026-10 (newest answer).

Spotted a mistake? Report it: tell us on X (@ElonsFAQ). Please mention Q37.

Related questions

More in AI risk, safety and valuesAll topics